Re: Alternate floating-point results under directed rounding
On Sun, 2008-11-09 at 23:13 -0800, John Pryce wrote:
> Similarly if a mixed point & interval computation goes close to
> overflow then - in genuine application code - that's dumb, and you
> should be _grateful_ for a mechanism, like returning Empty, that tells
> you so.
>
> Actually I say it should return NaI. Empty can be a genuine, correct
> result, whereas NaI is _always_ an indicator that something went
> wrong.
I don't understand the advantage of NaI as the result of a computation
such as 2.0*[0.25*HUGE,0.75*HUGE], instead of returning [0.5*HUGE,inf]
with overflow signaling.
NaI might be argued for 1.1*[0.95*HUGE,0.98*HUGE], but is that really
better than [inf,inf] with overflow signaling? When IEEE FPA produces
inf, we don't really know how much larger than HUGE the correct result
is: IEEE inf represents some unknown member of the set {x|x>HUGE}, so if
[inf,inf] is undesirable a better result might be the open interval
(HUGE:inf).
--
Van Snyder | What fraction of Americans believe
Van.Snyder@xxxxxxxxxxxx | Wrestling is real and NASA is fake?
Any alleged opinions are my own and have not been approved or
disapproved by JPL, CalTech, NASA, the President, or anybody else.